jeremyong/Selene

Support Visual Studio 2015

arvidsson opened this issue · 8 comments

I know that it was seemingly impossible to support VS2013. I'm not sure about the exact changes in VS2015 relating to supporting C++ standards, thus my question or wish: would it be possible to support VS2015?

I have VS 2015 Community and it works with the 140 Toolset. Also you need Lua 5.2 because VS doesn't seem to like the 5.3 libs.

VS 2015 works totally fine with Selene and lua 5.3.1
My proof:
https://github.com/bagobor/cpp2lua-buindings-battle

Feel free to be more precise (with exact error messages) 👍

Ok, I haven't tested it properly in VS2015 yet. I wasn't sure whether support was at all there, since I couldn't find an issue related to it, and I didn't want to waste time switching my project to VS2015.

by the way, you may notice that selene has worstest perfomance (compare to other luabinding libs)

@bagobor Yes you've already made an entire separate issue addressing that. Honestly, the benchmark just calls a single function that increments a number repeatedly. It's a poor benchmark for a number of reasons, and I've already addressed several. In addition, I've already pointed out an optimization that would make it perform identically to others listed in your benchmark. If you care about the performance to such a degree, just fix it instead of talking about it.

@jeremyong I'm not sure is it right place to discuss it.
As you may notice - there is 2 selene 'benchcases': a) selene b) selene_cached.
Second was added for your request to 'optimize' by caching lua state locally.

While dealing with lua - there is only only few most frequent operation - is to call some lua function, call some c-function from lua and get some output values. So my benchmark covered all that. (in a simple way for now). I will be glad to fix any bugs in my test scheme.

@bagobor If this isn't the right place to discuss it, don't bring it up.
I was referring to #94 which I created specifically to address the performance issue you pointed out.

As for "... there is only only few most frequent operation... " this is entirely subjective. I don't know how you could claim to know what the most typical use case of a user is. In fact, this faulty assumption is usually what the unfortunate majority of benchmark authors possess.

We are talking about layer between c++ and Lua, right?
There is no absolute truth, that's right.
What kind of benchmark will show the true power of Selene?

On Tuesday, September 15, 2015, Jeremy Ong notifications@github.com wrote:

@bagobor https://github.com/bagobor If this isn't the right place to
discuss it, don't bring it up.
I was referring to #94 #94
which I created specifically to address the performance issue you pointed
out.

As for "... there is only only few most frequent operation... " this is
entirely subjective. I don't know how you could claim to know what the most
typical use case of a user is. In fact, this faulty assumption is usually
what the unfortunate majority of benchmark authors possess.


Reply to this email directly or view it on GitHub
#99 (comment).